Development, Characterization, and Stability of Flavored Water Kefir: Impact of Fermentation and Storage

Abstract

The increasing demand for functional beverages sparked greater interest in health-promoting craft drinks, highlighting the need to optimize production parameters and assess their stability. This study aimed to develop, optimize, and characterize a grape juice-flavored naturally carbonated water kefir, evaluating its sensory qualities, physicochemical and microbiological stability. Fermentation conditions (F1) were optimized using Central Composite Rotational Design, leading to the selection of 24 h at 30 °C with (6.5% w/v) brown sugar, ensuring efficient pH reduction to safe levels. Sensory analysis selected grape juice as the flavoring agent, and a mixture design coupled with the desirability function determined the optimal formulation as 50% kefired water, 46.4% grape juice, and 3.6% water, resulting in high overall sensory desirability. During 42 days of refrigerated storage (4 °C), the beverage exhibited progressive sugar consumption from residual metabolic activity, a dynamic antioxidant profile characterized by increases in total phenolic compounds and FRAP activity, stability in ABTS activity, and decline in DPPH activity. Lactic acid bacteria counts remained stable during storage, while acetic acid bacteria and yeast populations decreased. Furthermore, pH (~3.30) and alcohol content (~1.86 °GL) remained stable, although the latter requires clear labeling in compliance with regulations for similar fermented beverages.

1. Introduction

The global demand for functional beverages has grown alongside increased interest in healthy eating, with similar trends observed in Brazil. In this scenario, foods with probiotic potential have gained prominence for their ability to promote intestinal health and strengthen the immune system [1], offering benefits that include restoring the microbiota and modulating the immune response [2].

Among functional beverage options, water kefir stands out as a traditional and versatile fermented product. Water kefir grains consist of a complex polysaccharide matrix, composed primarily of dextran and a levan fraction, harboring a symbiosis of microorganisms, including lactic acid bacteria (LAB), acetic acid bacteria (AAB), and yeast [3,4]. It has a gelatinous and translucent appearance, with irregular shapes and sizes. The microbial diversity and, consequently, the physicochemical composition of kefir can vary significantly depending on the origin of the grains, cultivation methods, raw materials, and processing technology [5,6]. This intricate metabolic process creates distinctive flavor and aroma profiles along with natural carbonation, resulting from the production of ethanol, lactic acid, and carbon dioxide. Furthermore, beneficial metabolites such as glycerol, mannitol, esters, and B vitamins are synthesized during the fermentation process [7].

The stability and sensory characteristics of water kefir are strongly influenced by nutrient types and concentrations, flavorings, and fermentation parameters [8,9]. Optimizing these proportions is key to achieving a stable final product that delivers health benefits. Water kefir has great potential to meet the growing demand for products suitable for specific audiences, such as vegans, lactose-intolerant individuals, and consumers with healthy eating habits. Furthermore, the presence of bioactive compounds with anti-inflammatory, antioxidant, and antimicrobial properties has been extensively investigated [10], which significantly contributes to its popularization.

In this context, systematically optimizing the formulation of fermented water kefir beverages is essential to ensure standardization, improve sensory and functional characteristics, and expand their market potential. The application of statistical experimental design tools allows for the efficient exploration of interactions between components and the identification of optimal proportions.

Water kefir is traditionally consumed as a plain fermented beverage (kefired water), but recent trends favor flavored and naturally carbonated kefir products to enhance sensory appeal and broaden consumer acceptance. To achieve this, a common approach involves a two-stage fermentation process. The first fermentation (F1) typically involves cultivating water kefir grains in a solution containing water and a carbohydrate source (such as brown sugar) in open containers to produce kefired water, a base liquid with a consistent microbial profile after the grains are filtered out. Subsequently, in the second fermentation (F2), this prepared kefired water is combined with flavoring agents, like fruit juice or vegetable extract, and fermented under closed conditions to promote the natural carbonation of the beverage by retaining the CO₂ produced from residual microbial activity [11].

Studies have explored various fruit and plant extracts for this purpose, such as Aronia melanocarpa juice and pomace by Esatbeyoglu et al. [6], strawberry juice by Araújo et al. [12], and fig extract by Laureys et al. [13]. However, little research has explored grape juice as a flavoring for water kefir, despite its high antioxidant content and strong consumer acceptance, making it a promising alternative for developing a novel water kefir product with enhanced functional and palatable attributes. In this context, this study systematically developed and characterized a grape juice-flavored water kefir through a comprehensive three-phase experimental approach. Phase 1 focused on optimizing the first fermentation (F1) conditions. Subsequently, Phase 2 concentrated on formulating the second fermentation (F2), guided by sensory evaluation to identify optimal proportions. Finally, Phase 3 rigorously assessed the stability of the optimized beverage under refrigerated storage, evaluating its physicochemical and microbiological properties over an extended period.

Thus, the objective of this study was to develop a fermented, carbonated, and flavored beverage from drinking water, brown sugar, fruit juice, and water kefir grains, and to evaluate its sensory, physicochemical, and microbiological aspects.

2. Materials and Methods

2.1. Study Overview

Water kefir can be consumed as a plain fermented beverage (kefired water) resulting from a single fermentation, or it can be further flavored and/or carbonated. Recent studies [3,7,12] have increasingly focused on flavored and carbonated water kefir. This study adopted a common methodology to achieve a naturally carbonated and flavored beverage, which involves a two-stage fermentation process:

First Fermentation (F1): Conducted in open containers with only water, a carbohydrate source (brown sugar or other), and water kefir grains. Grains are filtered out to yield “kefired water,” providing a base liquid with a consistent microbial profile, as it is challenging to standardize the grains by direct cell counts, due to their complex insoluble nature.

Second Fermentation (F2): The prepared kefired water is combined with fruit juice or vegetable extract for flavoring, and fermented under closed conditions to retain CO₂ from residual microbial activity, producing natural carbonation.

The experimental approach in this study was systematically divided into three main phases to ensure a comprehensive development and characterization of the flavored water kefir:

Phase 1: Optimization of the First Fermentation (F1). This phase focused on studying the fermentation kinetics of F1, varying the sugar concentration and fermentation temperature to achieve a safe pH and consistent kefired water quality for F1.

Phase 2: Formulation Development for the Second Fermentation (F2). Building on the optimized F1 conditions, this phase focused on identifying the proportions of kefir water, fruit juice, and additional water for the final beverage. Sensory evaluation was conducted to determine consumer preferences regarding different flavoring agents and their concentrations.

Phase 3: Stability Study of the Formulated Beverage. The beverage formulated with the parameters from previous phases underwent a thorough stability study under refrigerated storage, for assessment of its physicochemical and microbiological properties over an extended period.

2.2. Activation of Water Kefir Grains and Standard Fermentation of Fruit-Based Water Kefir

Water kefir grains (stock from the Laboratory of Unit Operations at UFES, Alegre, Brazil) were activated in a 6.5% (w/v) brown sugar solution at 30 °C in open containers. Grains were filtered and reused in fresh solution every 24 h for 7 days following Almeida [11] with modifications. Kefir grains were preserved under refrigeration at 4 °C for short-term storage or frozen at −10 °C for extended periods.

The beverage preparation comprised two consecutive fermentation stages: Fermentation 1 (F1) and Fermentation 2 (F2). In F1, kefir grains (5% m/v) were aerobically inoculated into a brown sugar solution (3–10% w/v) and incubated at 20–30 °C for 24 or 48 h. After removing the grains, the kefired water was flavored with whole grape juice (10–50%) and subjected to F2, which occurred anaerobically at 30 °C for 24 h. The finished beverages were then refrigerated at 4 °C for 24 h for analysis.

2.3. Study of Fermentation Kinetics: Influence of Temperature and Total Soluble Solids on Fermentation Kinetics (F1)

The kinetics of Fermentation 1 (F1) were investigated using a Central Composite Rotational Design (CCRD), varying temperature (20–30 °C) and soluble solids (3–10 °Brix) in 11 experimental units (

Table 1. Experimental planning matrix for Central Composite Rotational Design (CCRD).

Maximum Level (%)		Minimum Level (%)	Variables
30		20	Temperature (°C)
10		3	Soluble Solids (SS)
TSS (°BRIX)	Temperature (°C)	X2 (coded SS)	X1 (Coded temperature)	Essay
4.0251	21.5	−1	−1	1
8.9749	21.5	1	−1	2
4.0251	28.5	−1	1	3
8.9749	28.5	1	1	4
6.5	20	0	−1.4142	5
6.5	30	0	1.4142	6
3	25	−1.4142	0	7
10	25	1.4142	0	8
6.5	25	0	0	9
6.5	25	0	0	10
6.5	25	0	0	11

). The pH was monitored as a response variable at predefined intervals throughout 48 h of fermentation. Fermentations were carried out in 250 mL open beakers with 5% (m/v) kefir grains and temperature control. Ultimately, the grains were separated by filtration, and the fermented liquid (kefired water) was characterized by its pH, total titratable acidity, alcohol content, sugars, color, and counts of lactic acid bacteria and enterobacteria.

2.4. Selecting the Flavor of Fruit Juice Used to Flavor Water Kefir

Flavor selection was performed through sensory evaluation of the flavored formulations, using a Completely Randomized Design (CRD). Sensory attributes (color, aroma, flavor, carbonation, and overall impression) and purchase intention were compared. Data were analyzed using Analysis of Variance (ANOVA), and in cases of significant differences (p < 0.05), the means were compared using Tukey’s test.

2.5. Determination of the Proportion of Grape Juice in the Formulation of Water Kefir

Following flavor selection, a mixture design (

Table 2. Mixture Design Matrix (MDM) for choosing formulation proportions.

Maximum Level (%)			Minimum Level (%)			Variables
90			50			Kefired water
50			10			Fruit juice
40			0			Drinking water
Components			Pseudocomponents			Treatment
Water (%)	Juice (%)	Kefir (%)	x3	x2	x1
0	10	90	0	0	1	1
0	50	50	0	1	0	2
40	10	50	1	0	0	3
0	30	70	0	0.5	0.5	4
20	10	70	0.5	0	0.5	5
20	30	50	0.5	0.5	0	6
13.3333	23.3333	63.3333	1/3	1/3	1/3	7
6.6667	16.6667	76.6667	1/6	1/6	2/3	8
6.6667	36.6667	56.6667	1/6	2/3	1/6	9
26.6667	16.6667	56.6667	2/3	1/6	1/6	10

) tested different ratios of kefired water, grape juice, and water. Sensory attributes and purchase intent served as response variables.

Multivariate Optimization Analysis: Desirability Function

The formulation was defined using the desirability function [14] a methodology that allows the simultaneous optimization of multiple response variables to identify the ideal formulation conditions. This statistical approach transforms each evaluated response variable (such as sensory attributes and purchase intention) into an individual desirability (d_i), ranging from 0 (undesirable) to 1 (desirable), calculated according to Equation (1).

$$d_i=\left\{\begin{array}{c}\begin{array}{c}0,if {\widehat{y}}_i\le L_i\\ {\left({\displaystyle \frac{{\widehat{y}}_i-L_i}{T_i-L_i}}\right)}^{r_i},if L_i<{\widehat{y}}_i<T_i\end{array}\\ 1,if {\widehat{y}}_i\ge T_i\end{array}\right.$$

(1)

where

• $T_i$ is the target (optimal) value
• $L_i$ is the minimum value for responses that will be maximized

The overall desirability (D) was then calculated as the geometric mean of the individual desirability (d_i) (Equation (2)).

$$D={\left({\Pi }_{i=1}^N\right)}^{1/N}$$

(2)

Overall desirability (D), which ranges from 0 to 1, serves as an integrated measure of the formulation’s overall quality. The optimal condition is the one that maximizes D, indicating the point of greatest simultaneous desirability for all response variables.

2.6. Storage Stability Study

Based on the optimized treatment, a physical–chemical and microbiological stability study of the beverage was carried out under refrigeration (4 °C) for 42 days. Weekly analyses (0–6 weeks) monitored pH, alcohol content, soluble solids, sugars (glucose, fructose, and sucrose by HPLC), and counts of lactic acid bacteria, acetic acid bacteria, yeasts, and the presence of enterobacteria.

2.7. Analytical Methodology

2.7.1. Sensory Analysis

The sensory analysis was conducted with panels composed of participants recruited among students and employees from the Federal University of Espírito Santo (UFES), Alegre Campus, Brazil. They were randomly selected based on their regular consumption of fermented beverages and availability to participate in the study. All evaluators provided voluntary informed consent prior to participation. This study was approved by the Ethics Committee on Research in Human Beings of the Health Sciences Center of the Federal University of Espírito Santo (UFES), Brazil (opinion number 6.479.574).

Acceptance tests were performed using a 9-point structured hedonic scale for color, aroma, flavor, carbonation, and overall impression, where 1 corresponded to “dislike extremely” and 9 to “like extremely”. Purchase intention was evaluated using a 5-point scale. Samples (30 mL) were randomly coded with three-digit numbers and served in individual sensory booths at the Sensory Analysis and Food Technology Laboratories of the Center for Agricultural Sciences and Engineering (CCAE-UFES), Alegre-ES, Brazil. The evaluation followed standard sensory analysis procedures, adapted from Stone and Sidel [15].

2.7.2. Physicochemical Analysis and Phenolic Compound Content of the Fermented Beverage

The beverages were characterized by several physicochemical analyses, including total soluble solids (TSS), pH, ethanol concentration (in °GL—degrees Gay-Lussac), reducing and total sugars, and color. TSS was determined directly by digital refractometry (model DR-A1, ATAGO, Tokyo, Japan), with results expressed in °Brix. The pH was measured using a digital pH meter (model 826 pH mobile, Metrohm, Herisau, Switzerland). The quantification of reducing and total sugars followed the methodology of Maldonade et al. [16]. The concentrations of individual sugars—glucose, fructose, and sucrose—were determined by high-performance liquid chromatography (HPLC) using a Shimadzu system (Shimadzu Corp., Tokyo, Japan) equipped with an autosampler (SIL20AHT), column oven (CTO10ASVP), refractive index detector (RID10A) for sugar determination, and a system controller (CBM20A), with minor adaptations from the methodology of Bravim et al. [17]. Samples were centrifuged at 10,000× g at 4 °C for 15 min. The supernatant was collected, microfiltered through a 0.45 μm cellulose acetate filter, and transferred to 2 mL vials. A volume of 15 μL was injected into an Aminex HPX-87C column (300 cm × 7.8 mm) maintained at 55 °C. The chromatographic separation was achieved using a 5 mM sulfuric acid solution as the mobile phase, with a flow rate of 0.6 mL/min. Glucose, fructose, sucrose, mannitol, and glycerol were detected using a Refractive Index Detector (RID). Compounds were identified by comparing their retention times with those of pure standards and quantified using external calibration curves.

Color analysis was performed using a colorimeter (Spectrophotometer CM-5, Konica Minolta, Tokyo, Japan) to obtain the parameters L* (lightness), a* (green-red axis), and b* (blue-yellow axis). Alcohol content was measured with a portable meter (model ABV Valuer, Nova Creations Ltda., São José dos Pinhais, Brazil) after decarbonation of the sample in an ultrasonic bath (model UPL Ultrasonic system, CTA do Brasil Ltda, Campanha, Brazil) at 25 °C for 30 min, and dilution 1:3 (v/v), with the calculation adjusted by Equation (3):

$${Alcohol-content}_{\left(°GL\right)}=C.F_d$$

(3)

where C is the measured concentration and F_d is the applied dilution factor.

Additionally, the total phenolic compound content was determined using the Folin–Ciocâlteu methodology [12], with results expressed as mg of gallic acid equivalent/mL. Antioxidant capacity was evaluated using three assays: ABTS (2,29-azinobis-(3-ethylbenzothiazoline-6-sulfonic acid)), DPPH (1,1-diphenyl-2-picrylhydrazyl), and FRAP (iron reduction). The ABTS, FRAP and DPPH methods were performed as described by Araújo et al. [12], expressed as mg of Trolox equivalent/mL.

2.7.3. Microbiological Analysis of Fermented Beverage

The microbiological characterization of the fermented beverage focused on quantifying specific populations and detecting contaminants. Sample preparation for decimal serial dilutions was performed as described by Silva et al. [18].

Total lactic acid bacteria and acetic acid bacteria counts were determined by deep plating on specific media (MRS and GYC, respectively), with incubation under controlled conditions, following the methodology detailed by Silva et al. [18]. Yeast quantification was performed using a Neubauer chamber count with methylene blue staining, as adapted from Freitas et al. [19]. Finally, total enterobacteriaceae counts were performed using the pour plate method on VRBG agar, as described by Silva et al. [18].

2.8. Statistical Analysis

The statistical designs employed, such as the Central Composite Rotational Design (CCRD), the Completely Randomized Design (CRD), and the mixture design, were applied according to the nature of each experiment and have already been detailed in the previous methodological sections. For the analysis of quantitative data, nonlinear regression analysis was employed, with models adjusted using the least squares method, and the quality was assessed by the coefficient of determination (R²). In situations where no clear regression trend was observed, or for specific comparisons, Analysis of Variance (ANOVA) was used, with comparison of means by Tukey’s test (p < 0.05).

3. Results and Discussion

3.1. Influence of Temperature and Total Soluble Solids on Fermentation Kinetics (F1)

Figure 1 illustrates the pH variation curves over time for the various treatments, and

Table 3. Adjusted equations and coefficients of determination (R2) for different treatments tested.

Treatment	Adjusted Equations	R2
T1	$\widehat{pH}=5.5083e^{-0.005175t^{0.5714}}$	0.9954
T2	$\widehat{pH}=5.5475e^{-0.04072t^{0.6136}}$	0.9798
T3	$\widehat{pH}=5.5463e^{-0.06518t^{0.5865}}$	0.9776
T4	$\widehat{pH}=5.5526e^{-0.04762t^{0.6262}}$	0.9759
T5	$\widehat{pH}=5.5559e^{-0.04466t^{0.5476}}$	0.9957
T6	$\widehat{pH}=5.0384e^{-0.05586t^{0.5642}}$	0.9902
T7	$\widehat{pH}=4.6915e^{-0.03663t^{0.6143}}$	0.9866
T8	$\widehat{pH}=4.9953e^{-0.03617t^{0.6366}}$	0.9908
T9	$\widehat{pH}=5.0157e^{-0.04946t^{0.5637}}$	0.9851
T10	$\widehat{pH}=4.9660e^{-0.04562t^{0.5826}}$	0.9890
T11	$\widehat{pH}=5.0093e^{-0.04896t^{0.5683}}$	0.9831

presents the model parameters adjusted for fermentation kinetics.

The fermentation kinetics, assessed through pH variation over time, are shown in Figure 1. The results indicate that pH progressively decreased throughout the fermentation period, following a modified exponential model. This model describes a decay rate that changes over time rather than remaining constant—a behavior typical of biological processes such as fermentation. The fitted model is presented in Equation (4), and the corresponding parameter estimates are provided in Table 3.

$$\widehat{pH}=A{.e}^{-k{t}^B}$$

(4)

where A represents the initial or maximum pH value (i.e., the pH at time zero), e is the base of the natural logarithm, k denotes a decay rate constant, t corresponds to the fermentation time, and B is an exponent that directly influences the shape of the pH decay curve as a function of time.

The adjusted models demonstrated an excellent fit to the experimental data, with consistently high coefficients of determination (R²) for all treatments (Table 3). These values, ranging from 0.9759 to 0.9957, indicate that the models explained a significant portion of the observed variability in pH during fermentation, confirming the adequacy of the modified exponential model in describing the acidification kinetics under each experimental condition.

These results are similar to those reported by Zongo et al. [20], who, when studying the fermentation kinetics of a new palm sap-based kefir drink, observed a drop in the pH value from 6.53 to values around 4.00 after 48 h of fermentation at 22 °C. An exponential or modified exponential decay of pH as a function of time was also found by other authors, including Almeida [11].

The drop in pH values of the samples is expected and is primarily associated with the metabolic activity of lactic acid bacteria, acetic acid bacteria, and yeasts. These microorganisms not only produce acids but also generate ethanol, volatile compounds, and carbon dioxide [20].

The fermentation period is a key factor influencing the final characteristics of the product, with first fermentation times often reported ranging from 24 to 48 h in the literature [11,13]. To understand the effect of process time on pH for the different treatments, the pH was predicted after 24 and 48 h using the 11 modified exponential models fitted to the kinetic data presented in

Table 4. pH value after 24 and 48 h of fermentation.

Treatment	Initial pH (A)	pH (24 h)	pH (48 h)
T1	5.51	4.01	3.43
T2	5.55	4.17	3.58
T3	5.55	3.64	2.95
T4	5.55	3.92	3.24
T5	5.56	4.31	3.83
T6	5.04	3.60	3.07
T7	4.69	3.62	3.16
T8	5.00	3.80	3.26
T9	5.02	3.73	3.24
T10	4.97	3.71	3.21
T11	5.01	3.72	3.22

.

The initial pH of each treatment—corresponding to the pH at the start of fermentation (when time t = 0)—is expressed by parameter A of the modified exponential model. As shown in Table 4, the initial pH ranged from 4.69 (Treatment T7) to 5.56 (Treatment T5), demonstrating different starting conditions between the experiments. As expected, pH decreased markedly during fermentation. After 24 h, the pH values predicted by the models ranged from 3.60 (Treatment T6) to 4.31 (Treatment T5). This acidification trend continued through 48 h, with pH values ranging between 2.95 (Treatment T3) and 3.85 (Treatment T5).

Reducing the pH during fermentation is a crucial factor for ensuring the microbiological safety and preservation of fermented beverages. Reaching and maintaining a pH below 4.5 is a critical threshold, as it creates an environment unfavorable for the growth of most pathogenic microorganisms, especially sporulating microorganisms, which are unable to germinate and produce toxins in such acidic conditions [21].

Considering the acidification rate and lowest final pH value, the three best treatments would be T3, T6, and T7. Treatment T6 demonstrated strong performance (R² of 0.9902)—higher than that of T3 (R² of 0.9776) and T7 (R² of 0.9866)—presenting a slight advantage over the others. Although Treatment T3 presented the highest decay constant (k = 0.06518), indicating a potentially faster acidification rate, its significantly higher initial pH (5.55) resulted in a pH of 3.64 after 24 h, still higher than the pH of 3.60 achieved by T6 during the same period.

This demonstrates that, starting from a lower initial pH (5.04), T6 can reach the safe pH range efficiently and at an early stage. Within 48 h, the pH of T6 stabilizes at 3.07, a value very close to the lowest pH observed (T3 at 2.95 and T7 at 3.16). The combination of a rapid and effective pH reduction during the initial critical period, combined with an R² of 0.9902 that confirms the high accuracy of the model, led to the selection of T6 for subsequent experiments. Thus, the experiment continued using 6.5% (w/v) brown sugar and a fermentation temperature of 30 °C. This condition is within the range of values reported in the literature. Zannini [9] used 7.5% sugar and a fermentation temperature of 26 °C; Almeida [11] established conditions of 5% water kefir grains, 10% added brown sugar, and 25 °C; Arapović [22] used 5.56% sugar and 22 °C. Although no standardized conditions are established, the values proposed in the present study are consistent with those reported in previous works.

3.1.1. Physicochemical Characterization of the Fermented Beverage After 24 and 48 h

After selecting Treatment T6, the next step involved determining the ideal time for the first fermentation. However, this decision could not be based solely on pH values, as consumer acceptance is a critical factor for commercial success and product viability. Considering the importance of sensory analysis in guiding consumer preferences, a sensory study was conducted to compare the 24 and 48 h fermentation times in F1. To ensure comparability of results, the second fermentation (F2) and the amount of juice used for flavoring were standardized as described in the methodology. The kefired water obtained after 24 and 48 h of F1 (characterized in

Table 5. Physical, physicochemical, and microbiological parameters of the fermented beverage after 24 and 48 h of fermentation.

Physicochemical Parameters	24 h	48 h
Total soluble solids (°Brix)	7.77 a	7.63 a
pH	3.56 a	3.65 a
Alcohol content (°GL)	0.73 a	0.75 a
Total sugars (%)	1.80 a	1.30 b
Reducing sugars (%)	1.4 a	1.0 b
Lactic acid bacteria (log UFC)	7.13 a	7.44 b
Enterobacteria	Absence	Absence
Color L*	4.15 a	2.30 b
Color a*	12.87 a	7.27 b
Color b*	7.01 a	3.89 b

Note: Data expressed as mean (n = 3). Means followed by different letters on the same line differ statistically by Tukey’s test (p < 0.05).

) then underwent the standardized F2 process before being subjected to sensory evaluation.

Table 5 presents the results of the physical–chemical and sensory analyses of the fermented beverage at 24 and 48 h of fermentation.

Table 5 presents the physicochemical parameters of water kefir during the first fermentation (F1) at 24 and 48 h. The data confirm the expected sugar consumption, with a decrease from 6.5 (w/v) of added sugar at time zero to 1.80% of total sugars (24 h) and 1.30% (48 h). The reduction in sugar content may be related to the biosynthesis of other components, such as the polysaccharides that constitute kefir grains. It should be noted that fermentation continues during this period, albeit at a very low rate, which does not justify extending the fermentation time by an additional 24 h. Nevertheless, further studies are required to confirm this hypothesis.

No significant differences were observed for the parameters of soluble solids (7.77 °Brix in 24 h and 7.63 °Brix in 48 h), pH (3.66 in 24 h and 3.65 in 48 h), and alcohol content (0.73 °GL in 24 h and 0.75 °GL in 48 h). These results suggest that extending the primary fermentation (F1) from 24 to 48 h did not lead to meaningful changes in these parameters.

It is important to note that the pH values observed in Table 5 were slightly different from those calculated in Table 4. This discrepancy arises because the values in Table 4 were estimated using mathematical models, which, despite showing a good fit, cannot achieve perfect accuracy. In contrast, the values in Table 5 are based on actual measurements. Additionally, since kefir production is based on a wild culture (kefir grains), variations between different batches are common. Despite these differences, the observed values remain within an acceptable range and do not impact the conclusion of this research.

Regarding the alcohol content of the beverage, there were no significant differences between the 24 and 48 h fermentation times (p > 0.05), with an average value of 0.74% (v/v). According to Brazilian legislation, beverages containing less than 0.5% (v/v) alcohol are classified as non-alcoholic, as established by Decree No. 6871/2009 [23].

Although there is no specific legislation for water kefir in Brazil, and the Normative Instruction for milk kefir [24] does not define standards for alcohol content, the analogy with kombucha is the best option for labeling and classification purposes. Kombucha, a fermented beverage that shares process similarities with water kefir by using SCOBY inoculum containing lactic acid bacteria, acetic acid bacteria, and yeast, is regulated in Brazil by Normative Instruction No. 41/2019 [25]. This regulation establishes that if the alcohol content of kombucha exceeds 0.5% (v/v), then it must be declared on the label.

Given the values found in this study (an average of 0.74%), which exceed the 0.5% (v/v) threshold, water kefir should therefore have its alcohol content declared on the label, following the precedent established for fermented products with a similar profile. Maintaining the alcohol content below 0.5% (v/v) is, in fact, a considerable challenge for producers of fermented beverages marketed as non-alcoholic, such as kefir, kombucha, tepache, among others. Tran et al. [26] define kombucha as a low-alcohol fermented beverage, illustrating that such beverages inherently produce small but measurable amounts of alcohol.

Ethanol production is an inherent characteristic of water kefir fermentation, driven primarily by the activity of yeasts present in the grains. The scientific literature presents a wide range of alcohol contents for water kefir, ranging from trace amounts to over 2% (v/v), depending on factors such as the predominant yeast strain, initial sugar concentration, and fermentation time/temperature [27].

Regarding the total sugar content, a slight reduction was observed (from 1.80% to 1.30%), indicating ongoing fermentation activity. This continuity is corroborated by the increase in the lactic acid bacteria population. The sugar consumption profile, characterized by a preference for glucose and the subsequent formation of ethanol and organic acids such as lactic and acetic acids, reflects the classic metabolic pattern of water kefir fermentations, as described by Laureys [13]. This pattern is characterized by a stable microbial community, where lactic acid bacteria (LAB) consistently outnumber yeasts (typically a 2:1 to 10:1 ratio), even though yeasts are primarily responsible for major metabolite production like ethanol. The initial preference for glucose consumption over fructose is also a key feature.

Furthermore, changes in the color parameters (L*, a*, b*) revealed that the drink became considerably darker and lost its reddish and yellowish hues during this period. While these data alone do not allow us to determine the optimal fermentation time, they suggest little difference between 24 and 48 h. Therefore, a sensory analysis was conducted to provide a more definitive assessment.

3.1.2. Sensory Evaluation of the Fermented Beverage After 24 and 48 h

Table 6. Result of sensory evaluation of fermented beverage at 24 and 48 h of fermentation.

48 h	24 h	Sensory Attributes
6.13 ± 1.7 a	5.98 ± 1.5 a	Color
5.68 ± 1.6 a	5.70 ± 1.8 a	Aroma
6.27 ± 2.0 a	6.25 ± 1.9 a	Flavor
6.90 ± 1.8 a	7.01 ± 1.6 a	Carbonation
6.52 ± 1.4 a	6.45 ± 1.3 a	Overall impression

Different letters on the same line indicate significant difference by Student’s t-test (p < 0.05).

presents the average scores from the sensory analysis of the beverage after 24 and 48 h in the first fermentation (F1). In both cases, 50% whole grape juice was added, followed by a second fermentation of 24 h at 30 °C to enhance flavor and carbonation. Thus, the only difference between the two treatments is the duration of the first fermentation (F1). Prior to sensory evaluation, the microbiological evaluation confirmed the absence of Enterobacteriaceae, ensuring the product’s safety and suitability for sensory testing.

The sensory evaluation of the fermented beverages, summarized in Table 6, revealed a relevant finding for process optimization. No statistically significant differences (p > 0.05) were observed between the 24 and 48 h fermentation times for color, aroma, flavor, carbonation, overall impression, or purchase intention. Acceptance scores for hedonic attributes ranged from 5.68 (Aroma) to 7.01 (Carbonation) on a scale of 1 to 9, indicating satisfactory acceptance in these aspects. The carbonation attribute received the highest score, indicating that the second fermentation under closed conditions produced an effervescent beverage aligned with consumer preferences.

Given the sensory equivalence and the absence of significant differences between fermentation times for all evaluated attributes, a 24 h initial fermentation (F1) was selected for subsequent experiments. From a process optimization standpoint, this choice offers clear advantages: shorter processing times reduce operating costs, such as energy consumption and equipment usage, while also accelerating the production cycle, enabling higher output in less time.

3.2. Selecting the Flavor of Fruit Juice Used to Flavor Water Kefir

Based on the previously optimized operating conditions for the first fermentation (6.5% w/v) sugar and 30 °C), this study proceeded to the flavoring stage. To this end, the kefired water obtained in the first fermentation was added with 50% of different fruit juices (berries, pineapple, and grape), followed by a second fermentation conducted in a closed system at room temperature. This berry blend consisted of strawberries, raspberries, and blueberries. A sensory analysis was performed to evaluate the acceptance and sensory perception of these flavored formulations, and the results are presented in

Table 7. Average sensory scores for different attributes evaluated and different flavors.

Attributes	Berries	Pineapple	Grape
Color	6.72 ± 2.0 ab	6.73 ± 1.5 a	6.18 ± 1.8 b
Aroma	5.86 ± 1.8 a	5.64 ± 1.9 a	5.65 ± 1.8 a
Flavor	6.28 ± 1.9 a	6.06 ± 2.0 a	6.21 ± 2.0 a
Carbonation	6.17 ± 2.0 b	6.54 ± 2.0 ab	6.89 ± 1.8 a
Overall impression	6.33 ± 1.6 a	6.14 ± 1.7 a	6.50 ± 1.6 a
Microbiology
1—Lactic acid bacteria	1.04 × 107 ± 0.19 a	5.9 × 107 ± 0.13 a	1.4 × 107 ± 0.17 a
2—Enterobacteriaceae	Absent	Absent	Absent

Means followed by the same letter, in the same line, do not differ from each other according to Tukey’s test (p > 0.05).

.

The comparative sensory evaluation of fermented beverages flavored with grape, berries, and pineapple juices, whose average scores are presented in Table 7, revealed both similarities and distinctions in their sensory profiles.

For the carbonation attribute, grape juice obtained the highest average (6.89), showing a statistically significant difference compared to the berries juice (p < 0.05), while not differing from pineapple juice. This superior carbonation may be related to the composition of grape juice, specifically its balance between acidity and sweetness, which favors the production and retention of carbon dioxide during the second fermentation.

Regarding color, grape juice (6.18) obtained a statistically lower score (p < 0.05) when compared to pineapple juice (6.73). Berries juice (6.71) did not differ significantly from either of the other flavors for this attribute.

For the other attributes, aroma, flavor, and overall impression, no statistically significant differences were observed among the three flavors evaluated (p > 0.05). The average scores for these attributes ranged from 5.64 (pineapple aroma) to 6.50 (overall grape impression), indicating moderate to good acceptance by the evaluators for these characteristics.

The comparative sensory analysis of beverages flavored with different fruit juices (berries, pineapple, and grape) revealed more similarities than differences in sensory acceptance among the various flavors tested. Microbiological analysis (LAB) also showed no significant differences across the samples.

Given this sensory equivalence and considering that preliminary tests were conducted with grape juice, which also offers competitive cost and greater market availability, the grape-flavored formulation was selected for further development of the beverage.

3.3. Determination of the Proportion of Grape Juice in the Formulation of Water Kefir

To optimize the beverage formulation, a blend design was developed using three components: kefired water, grape juice, and drinking water.

Table 8. Regression models adjusted for results of sensory analysis according to proportion of kefired water, grape juice, and drinking water.

Variable	Adjusted model	R2
Color	$\widehat{y}=5.4668x_1+7.5195x_2+5.2580x_3-1.5829x_1{x}_2+0.9558x_1{x}_3+4.2218x_2{x}_3-3.1895x_1{x}_2{x}_3$	0.9038
Aroma	$\widehat{y}=5.0173x_1+6.1609x_2+4.9522x_3-1.3473x_1{x}_2+0.8278x_1{x}_3+1.5348x_2{x}_3-2.3268x_1{x}_2{x}_3$	0.8433
Flavor	$\widehat{y}=4.9719x_1+7.1302x_2+3.8271x_3-2.2156x_1{x}_2+2.1413x_1{x}_3+3.4948x_2{x}_3-4.6928x_1{x}_2{x}_3$	0.9940
Carbonation	$\widehat{y}=5.9737x_1+6.9277x_2+5.2801x_3-1.8021x_1{x}_2+0.1373x_1{x}_3+1.6996x_2{x}_3-3.0327x_1{x}_2{x}_3$	0.9580
Overall impression	$\widehat{y}=5.2089x_1+7.2223x_2+4.3660x_3-2.4462x_1{x}_2+1.8164x_1{x}_3+3.0285x_2{x}_3-4.1307x_1{x}_2{x}_3$	0.9881
Purchase intention	$\widehat{y}=3.5854x_1+2.2352x_2+4.1372x_3+1.2951x_1{x}_2-0.9651x_1{x}_3-2.1924x_2{x}_3+4.1731x_1{x}_2{x}_3$	0.9969

Y is response for evaluated sensory attribute, x1 is amount of kefired water (%), X2 is amount of grape juice (%), and X3 is amount of water (%) added.

presents the models obtained for the sensory analysis responses.

Mathematical models were fitted to describe the sensory analysis results for the different treatments. The regression models demonstrated an excellent fit to the experimental data, with coefficients of determination (R²) ranging from 0.8433 to 0.9969. Using these models, the graphs shown in Figure 2 were obtained.

The multivariate optimization of the beverage formulation, conducted using the Derringer and Suich [14] desirability function, resulted in a maximum overall desirability of 0.9917 (Figure 3). This value, extremely close to ideality (1.0), represents the equilibrium point where multiple sensory attributes (such as flavor and carbonation) and purchase intention were simultaneously maximized.

The optimization of the mixture design, as illustrated in Figure 3, indicated a maximum overall desirability of 0.9917. This value, very close to the ideal desirability (1.0), was achieved with the pseudocomponent ratio of x₁ = 0.00, x₂ = 0.91, and x₃ = 0.09. In terms of real components, this combination is equivalent to 50% kefired water, 46.4% grape juice, and 3.6% water. This formulation represents the model’s optimal balance, maximizing all sensory attributes alongside purchase intention simultaneously. Based on this optimized composition, the final product was prepared and subsequently assessed for physical and chemical stability during storage.

Based on the data obtained for an optimized formulation (50% kefired water, 46.4% grape juice, and 3.6% drinking water), an acceptance test was conducted, yielding a satisfactory sensory evaluation of the analyzed attributes and validating the effectiveness of the optimization. Color received an average score of 5 (“indifferent”), suggesting potential for visual improvement. However, aroma, flavor, and carbonation received scores between 6 and 7, corresponding to “slightly liked” and “moderately liked,” respectively. The overall impression was also rated as “moderately liked.” Purchase intention varied, with a predominance of responses between “probably would buy” and “maybe would buy,” reflecting intermediate to positive consumer acceptance, which highlights the opportunity to improve the product’s color to further increase visual desirability.

Based on the data obtained, a standard formulation was established, as presented in

Table 9. Operational parameters and formulation used in production of naturally grape-flavored water kefir.

Parameters	F1	F2
Type of fermentation	Open	Closed
Fermentation time	24 h	24 h
Temperature	30 °C	30 °C
Added sugar (w/v)	6.5%	na
Quantity of grains (w/v)	5.0%	na
Kefired water	na	50%
Grape juice	na	46.40%
Drinking water	Complete 100%	3.60%

na: not applicable.

.

3.4. Stability of Water Kefir During Storage

3.4.1. Variation in Sugar Content of Water Kefir During Refrigerated Storage

Using the optimized data, grape-flavored water kefir was prepared and its stability monitored for 42 days of refrigerated storage at 4 °C, with samples collected at days 0, 7, 14, 28, 35, and 42. Figure 4 illustrates the variation in sugar content throughout this period, along with the fitted models and their corresponding coefficients of determination (R²).

A progressive and significant reduction in the concentration of all sugars is observed as storage time increases, indicating ongoing residual metabolic activity; that is, fermentation continues, albeit at a slower rate. It is essential to note that the curves fitted by the model exhibit an excellent fit to the experimental data, thereby validating the representativeness of the observed trends.

Individual sugar analysis corroborates microbial selectivity. The rapid decrease in sucrose can be attributed to the action of hydrolytic enzymes (such as invertases) produced by microorganisms that hydrolyze sucrose into glucose and fructose, making them available for consumption [28]. Glucose, in turn, was shown to be the monosaccharide preferentially consumed, showing the steepest decrease. This preference for glucose over fructose is a metabolic pattern frequently observed in water kefir fermentations and other microbial systems [8], where glucose is metabolized more efficiently, while fructose tends to persist in higher concentrations or be consumed more slowly, as also evidenced in the results of this study.

The total sugars curve, which represents the sum of the individual concentrations, reflects the general consumption trend, decreasing from an initial concentration of approximately 12.5 g/100 mL to approximately 5.0 g/100 mL after 40 days. Most of this loss occurred in the first 10–20 days of storage, with the degradation rate decreasing afterward, consistent with first-order kinetics, where the reaction rate is proportional to the substrate concentration.

These results suggest that, despite the completion of primary and secondary fermentation, sugar utilization continues during refrigerated storage, which can directly impact the perception of sweetness and the sensory characteristics of the beverage over time.

3.4.2. Variation in Antioxidant Capacity and Phenolic Compound Content of Water Kefir During Refrigerated Storage

For ABTS values, there was no significant variation over storage time (p > 0.05). In contrast, antioxidant capacity measured by DPPH and FRAP assays, along with phenolic compound content, showed significant variations (p < 0.05), as illustrated in Figure 5.

Figure 5 comprehensively illustrates the variation in total phenolic compound (TPC) content and antioxidant activity over the 42 days of refrigerated storage (4 °C) of the fermented beverage. The graph also displays the adjusted mathematical models, which, as shown in the figure, present an excellent fit to the experimental data.

The increase in total phenolic compounds and FRAP activity is a significant and unexpected finding. This rise is likely linked to the fermentation processes of microorganisms with potential probiotic properties and the transformations that occur in the matrix during storage. Grape juice fermentation with probiotics, as in this beverage, is known to increase total phenolic content. Probiotic bacteria can degrade tannins and produce compounds with a high content of free hydroxyl groups, which enhances phenolic content and antioxidant activity [29]. This biotransformation releases bound phenolics, increasing their availability and reactivity in assays like Folin–Ciocâlteu (TPC) and FRAP, which measures ferric ion reduction capacity [30].

Despite the increase in TPC and FRAP, the decrease in free radical scavenging activity (DPPH) is a behavior that, although seemingly paradoxical, is explainable by the complexity of the antioxidant profile. It is possible that the original antioxidant compounds or specific forms with high affinity for the DPPH radical are more susceptible to oxidative degradation or polymerization reactions during storage [31]. While new phenolic compounds may form, raising TPC and FRAP values, these might be less effective in the DPPH assay but exhibit greater electron-donating capacity, thus showing higher activity in the FRAP assay. The dynamics of phenolic compounds are influenced by several chemical reactions that can occur during aging, where some phenolics degrade while others remain stable, such as gallic acid [32].

The stability of ABTS activity suggests that the beverage maintains a general electron-donating/radical-scavenging capacity, perhaps through less specific compounds or mechanisms that are not as impacted by the transformations observed in DPPH. The diversity of microorganisms present [8,28] and the composition of the grape matrix itself contribute to this complexity and the multifaceted behavior of antioxidant capacity [30].

Thus, the results indicate a favorable evolution of the beverage’s antioxidant functionality, with the fermentation and storage processes promoting the formation or release of compounds that increase the total phenolic content and antioxidant reducing power. The particularity of the DPPH reduction highlights that different assays evaluate different aspects of antioxidant activity, and that the beverage undergoes a rearrangement of its bioactive profile, which is overall beneficial to health [33].

3.4.3. Variation in Water Kefir Microorganism Count During Refrigerated Storage

Assessing the population dynamics of microorganisms during refrigerated storage is crucial for understanding the stability and maintenance of the functional characteristics of fermented beverages such as water kefir. It is essential to note that, at zero storage time, the beverage has already undergone two fermentation phases: 24 h for the first fermentation and, subsequently, another 24 h for the second fermentation. This process ensures that microbial populations are already at or near their maximum growth rate, typically in the stationary or initial decline phase, rather than in the exponential growth phase.

Regarding the lactic acid bacteria (LAB) count, the results of this study indicate that there was no significant variation over the refrigerated storage time (p > 0.05). The average count of these bacteria remained at 3.78 × 10⁷ CFU/mL in the beverage.

This stability of the LAB population is a positive indicator, as maintaining high LAB counts is a characteristic not only associated with the health benefits of probiotic products like kefir, whose health-promoting potential, driven by bioactive compounds and beneficial microorganisms, has been extensively reviewed [34], but also observed in other traditional fermented foods and food matrices. Comparing this result with the literature, it is observed that the count obtained is in line with the expected levels for fermented water kefir products. For instance, Gökırmaklı et al. [35] reported Lactobacillus sp. counts in water kefir ranging from 5.18 to 7.71 log CFU/mL in fig-based media, and L. acidophilus at similar levels. The mean count of 7.58 log CFU/mL falls within the upper range or is slightly above the values reported in the literature for lactic acid bacteria in water kefir-related products, thereby reinforcing both the quality and probiotic potential of the developed beverage. Similarly, Almeida [11] observed lactic acid bacteria counts between 5.42 and 6.71 log CFU/mL during the fermentation stage and 7.17 ± 0.43 log CFU/g in the feed solution prior to drying, further confirming the high microbial viability achieved in this work

In contrast to lactic acid bacteria, acetic acid bacteria counts showed a significant decline over the storage period. This behavior is illustrated in Figure 6, where the population decay and the fit of an exponential model for this variation are observed.

The observed decrease in acetic acid bacteria (AAB) counts during refrigerated storage, as shown in Figure 6, reflects the characteristic metabolism of these microorganisms under the storage conditions. Acetic acid bacteria, such as those of the genera Acetobacter and Gluconobacter, are typically aerobic, relying on oxygen to carry out their main metabolic activities, such as the oxidation of ethanol to acetic acid and of sugars to gluconic acid [28,35].

Considering that the beverage, after both fermentations, was stored under refrigerated conditions in hermetically sealed bottles, oxygen availability became a critical limiting factor. The reduction in available oxygen in the environment leads to a decrease in the metabolic activity of AAB, and consequently, their inactivation and gradual death, resulting in the observed population decline, modeled by an exponential curve. This is a striking difference compared to lactic acid bacteria, which have anaerobic or facultatively anaerobic metabolism and are therefore less affected by the absence of oxygen.

Although direct comparative studies of AAB counts over storage time in water kefir are scarce, the initial presence of these bacteria in fermented beverages is well documented. Almeida [11], for example, detected acetic acid bacteria in their water kefir samples, with counts ranging from 5.04 to 5.75 log CFU/g, representing the starting point for the AAB population at zero storage time. The decline from these initial levels, as evidenced in the present study, directly reflects the impact of post-fermentation storage conditions.

The decrease in the AAB population carries both sensory and chemical implications for the beverage. Acetic acid is one of the main contributors to the characteristic flavor and aroma of water kefir, imparting a pungent acidity [35]. If this characteristic acetic acid level is too high, it can cause sensory rejection of the product. Additionally, the inactivity of AAB under anoxic conditions helps stabilize the beverage’s alcohol content, since ethanol oxidation, one of its primary functions, is inhibited. The decrease in AAB in a closed refrigerated environment can be seen as a natural control mechanism that prevents over-acidification and limits the formation of undesirable volatile compounds over shelf life.

The yeast population dynamics during refrigerated storage of water kefir exhibit a distinct pattern, differing both from both the stability of lactic acid bacteria and the decline of acetic acid bacteria. Initially, the results suggest that, following the two fermentation stages, the yeast population enters a stationary phase, which is later followed by a period of decline (Figure 6B).

According to the process history, the beverage at storage time zero has already completed two fermentation stages (24 h of F1 and 24 h of F2), which suggests that the yeast population has already reached or is close to its maximum growth peak. In other words, the yeast entered storage in a plateau phase, where the growth rate balances with the cell death rate, or a transitional phase toward decline, reflecting post-fermentation metabolic and environmental conditions.

In the early stationary phase, yeasts, having consumed the most easily assimilated substrates and produced metabolites such as ethanol and organic acids during fermentations, reach maximum population density. At this point, nutrient limitation and the accumulation of byproducts may begin to inhibit further growth. Gökırmaklı et al. [35] observed that yeast counts in sugar- and fig-based water kefir peaked after 24 h of fermentation, ranging from 5.89 to 6.16 log CFU/mL, before a slight decline between 24 and 48 h. Almeida [11] also reported yeast counts ranging from 5.03 to 5.65 log CFU/g.

Subsequently, the yeast population enters a decline phase. This phenomenon is driven by several factors, including nutrient depletion and low storage temperatures, accumulation of inhibitory metabolites, and cellular lysis and senescence, among others [28,36].

3.4.4. Variation of pH Value, Soluble Solids and Alcohol Content over Time During Refrigerated Storage

During the 42-day storage period, there was no significant difference (p > 0.05) in relation to the pH value and alcohol content. The average pH value was 3.30, and the average alcohol content was 1.86% °GL.

pH stability is an important indicator of both the quality and safety of fermented beverages over time, reflecting residual metabolic activity and product preservation. In the present study, analysis of water kefir beverages during 42 days of refrigerated storage revealed no variations over time in either pH or alcohol content (p > 0.05). The average pH value for this period remained at 3.30.

Achieving and maintaining a low pH, such as 3.30, is crucial for ensuring the microbiological safety of fermented beverages, as it creates an unfavorable environment for the growth of most pathogenic microorganisms [21]. This value is well below the critical limit of 4.5, commonly accepted as a barrier to pathogen development, and is consistent with the value found in other stages of the experiment.

Laureys [8] reported pH values ranging from approximately 3.34 to 3.47 after 48 to 192 h of fermentation in different water kefir formulations. Therefore, the pH value of 3.30 found in the present study is in line with the lower values observed by these authors. On the other hand, Almeida [11], in a study on powdered water kefir, reported pH values ranging from 3.88 to 4.30 during the first fermentation, differing from the present results.

Beyond safety concerns, pH has a direct impact on the flavor profile. Lower pH values, such as 3.30, indicate higher acidity, resulting in a sourer taste [37]. This acidity intensity can influence overall acceptance; for example, kefir with a pH of 3.37 was considered less preferred compared to slightly higher pH options [36].

The pH stability during storage, along with a consistent alcohol content (1.86 °GL), aligns with the observed dynamics of microbial populations. The stable population of lactic acid bacteria, combined with declining acetic acid bacteria and yeasts (as discussed previously), results in reduced metabolic activity, preventing significant fluctuations in acid and ethanol production.

On the other hand, there was a significant variation (p < 0.05) in the total soluble solids during storage time, as illustrated in Figure 7.

This decrease in TSS directly reflects the consumption of sugars present in the beverage by residual microorganisms, as shown in Section 3.4.1 and Figure 4, where a reduction in sugar levels is observed. Since sugars make up most of the soluble solids in water kefir and serve as the primary substrate for microbial activity, this decline indicates ongoing, albeit slow, residual fermentation.

This trend is consistent with the specialized literature. Esatbeyoglu et al. [6] observed a reduction in TSS values and sugar conversion, accompanied by a decrease in total sugars, glucose, and fructose, when studying water kefir produced from aronia. Complementarily, Pendón et al. [38] explained the fundamental process of sucrose consumption by yeast (with hydrolysis into glucose and fructose) and the subsequent use of these monosaccharides by lactic and acetic bacteria during water kefir fermentation. Although slowed by refrigeration, this residual metabolic activity is primarily responsible for the decrease in sugars and, consequently, in soluble solids.

It is worth noting that, despite the reduction in soluble solids content and sugar consumption, there was no significant difference in the pH value or alcohol content of the beverage during refrigerated storage. This suggests that residual sugars were primarily used by microorganisms for maintenance metabolism and cell survival, as well as for processes that do not result in significant production of additional acids or ethanol. Beverages can form a buffer system, allowing the pH to remain relatively stable even with small acid production.

4. Conclusions

The results enabled the establishment of process and formulation parameters that contribute to the production of a naturally carbonated, grape juice-flavored water kefir beverage with desirable sensory and functional qualities. This study successfully met its objectives by optimizing the formulation and evaluating the physicochemical, microbiological, and sensory stability of the product during refrigerated storage. The data obtained from this study enabled the definition of suitable conditions for producing a beverage with relevant sensory and functional attributes.

Analysis of the first fermentation (F1) revealed that the pH acidification kinetics follow a modified exponential model, influenced by temperature and soluble solids content. The selection of Treatment T6 (6.5% w/v brown sugar at 30 °C) and the 24 h initial fermentation time were justified by the combination of efficient acidification, achieving safe pH values, and the absence of significant differences in sensory attributes between 24 and 48 h of fermentation.

In the flavoring stage, grape juice was chosen for its effective enhancement of the beverage composition, with sensory acceptance comparable to that of other flavors tested, and notable carbonation. Using mixture design and desirability function, the optimized formulation was determined as 50% kefired water, 46.4% grape juice, and 3.6% drinking water.

During refrigerated storage (4 °C for 42 days), a progressive reduction in sugar (sucrose, glucose, and fructose) levels was observed, demonstrating ongoing residual metabolic activity. The beverage’s antioxidant capacity demonstrated dynamic behavior, with increases in total phenolic compounds and FRAP activity, and stability in ABTS activity, while DPPH activity decreased. The lactic acid bacteria population remained stable at significant levels, reinforcing the beverage’s probiotic potential over time. In contrast, acetic acid bacteria and yeast counts declined, likely due to oxygen and nutrient limitations. The stability of pH and alcohol content during storage is an important indicator of microbiological safety and the maintenance of product quality.

Although water kefir is often characterized as a low-alcohol beverage, the average values of 1.86 °GL observed during storage indicate the need to declare the alcohol content on the product label. This requirement aligns with current regulations for similar fermented beverages. It is essential to provide adequate information to consumers with specific restrictions on alcohol consumption, such as nursing mothers and drivers.

The objective of this study was successfully achieved, with advances in knowledge about the production of flavored and naturally carbonated water kefir. However, gaps in knowledge were identified that warrant further research. It is recommended that subsequent studies explore strategies for reducing the beverage’s alcohol content, aiming to expand its applicability to niche markets with restrictions on alcohol consumption. Additionally, a deeper understanding of microbial dynamics and interactions with metabolites produced during storage is essential to better understand the product’s stability and functional characteristics.